In message <535dae1be8...@timil.com>
Tim Hill <t...@timil.com> wrote:
> In article <6719835d53.br...@helpful.demon.co.uk>, Bryan Hogan
>> because the OS JPEG routines, as used in SwiftJPEG for example, can
>> display it at any scale in about one second.
> One second? Wow. Is that to repaint the whole window? Should've bought me
> an Omega.
:-) A more accurate timing reveals it to be 1.2s
> This Iyonix is taking about 4 seconds with Geminus (6.5 without) to
> display from RAMFS (whatever the quality setting) in SwiftJPEG.
I guess it is partly due to the Omega having the video memory in main
RAM rather than having to go across the PCI bus like the Iyonix, and
also the improvements in RO Select (cached screen memory, better JPEG
routines).
> I was just trying to demonstrate that NetSurf's lack of speed is not unique
While I'm not surprised that NetSurf's routines are slower than those
in RISC OS, a whole order of magnitude slower is surprising!
> That NetSurf doesn't seem to cache for very long a smaller version of the
> image that is painted into the window may seem odd
Is it a bug that NetSurf throws away the memory used by an image that
is still on screen? That results in multiple slow redraws rather than
just the initial ones.
Bryan.
--
RISC OS User Group Of London - http://www.rougol.jellybaby.net/
RISC OS London Show - http://www.riscoslondonshow.co.uk/
